In order to protect an audiovisual sequence against piracy, it is known to watermark the audiovisual sequence, visibly or invisibly, so as to identify the owner of a pirate copy.
It is also known to encode the audiovisual sequence during a transmission in order to prevent the illicit reading of the sequence.
It is certainly possible to combine these two protection processes, by encoding the audiovisual sequence at an encryption module on the transmitter side, and by watermarking the sequence after the decryption at a watermarking module on the receiver side.
However, such a process would not prevent a fraudster from recovering the sequence at the exit of the decryption module before passing to the level of the watermarking module.
Such a fraudster could then freely use the not watermarked audiovisual sequence, if he succeeded in deciphering the sequence.
In order to solve this general problem, a process for the distribution of a marked audiovisual sequence from a nominal audiovisual sequence towards a receiver equipment is known, said nominal audiovisual sequence having a nominal content, the process comprising steps in which:                a first mark is generated capable of being applied to the initial content at marking positions;        a modified stream is generated having a modified content different from said nominal content;        a complementary stream is generated comprising calculated complementary data so that it is possible to reconstitute an altered audiovisual sequence from said modified stream according to said complementary data,        said modified stream and said complementary stream are transmitted towards said receiver equipment.        
Such a process is known from the application WO 2004/062281.
In one embodiment of this application, the complementary stream includes marking instructions in order to insert an invisible and personalized mark into the marked audiovisual sequence.
In a preferred embodiment of the application WO 2004/062281, these instructions in particular allow to invert the LSB of certain visual coefficients, such as the DC coefficients.
The absence or not of an inversion on a LSB will allow, during an identification step, to determine the first mark inserted into the audiovisual sequence.
These instructions being inserted in the complementary stream, a fraudster could not have access to the audiovisual content before the application of the marking instructions, and thus before the insertion of the first personalized mark.
However, such a process does not allow an effective protection during a broadcast to receiver facilities.
Indeed, in broadcasting, the data transmitted to different receivers are identical for all the receivers.
Hence, if one applied the process described in the embodiment of the application WO 2004/062281 previously described, the marking instructions contained in the complementary stream would be identical for all the receivers and thus it would not be possible to distinguish pirate copies.
At the light of the aforementioned document, a problem that the invention sets out to solve is to allow a personalized marking of the audiovisual sequence distributed by broadcasting.
It could therefore be helpful to provide a process such as previously described in which:                one calculates said complementary data according to said marking positions;        one generates, on the receiver equipment, a second mark different from the first mark and        one reconstitutes said audiovisual sequence marked by the second mark at said marking positions from said modified stream according to said complementary data.        